Funny thing is we are all different. If they were to test again they would likely get different results. These researchers can not account for everything in their lives that can make said changes. It may be as simple as now we have more of a disconnect physically in society. IE smart phones & the internet. It could be attributed to many other metrics in our lives.
I can test the population of the DMS for a trait but I wouldn’t claim the correlation of that trait is the same for the global population. DMS is not a representative example of global population. It is like saying my bike is blue, thus I would be accurate in claiming all bikes are probably blue.
Without doing some statistical power analyses, no you can’t make those statements.
But a retrospective study of over 200,000 US adolescents with ADHD being treated with stimulants isn’t being extrapolated to the entire world. It’s of little utility to an octogenarian female in Siberia, for example. It’s up to the peer reviewers to decide whether the statistical methodology is solid.
Saying a sample size is too small without doing any math is just guessing. I for one have testified in court as to the validity of a statistical sample of only 30 healthcare claims and extrapolated the results of the review of my sample to a universe of several hundred thousand claims. My statistics showed that the sample was valid. And the healthcare provider who was committing fraud is now in jail.
1 Like
From the perspective of CORRELATIVE data, sure. For example if that study had been done when I was in the field, and I had a client with ADHD, and said client was on Adderall, then I might keep in the back of my mind that there is a very slight correlation that s/he has a higher propensity toward having a psychotic break. Would I put a ton of faith into it as a fact? hell no, each client was their own individual and you cant measure it like that. But what those kinds of studies do for practicing therapists in the field is maybe give some additional connections to things we didn’t know before, still cant call causal fact, but will file away in the back of our brains as possibilities for when working with said clients. Every persons brain chemistry is different so even with causal data to draw from, there’s no way to know how an individual will handle a new med.
and James just added additional flavor to the application of said analysis to populations way better than I could have stated so no need to say more there.
1 Like
This explains the crux of my issue. When you apply the data to the correct group, rather than the population of the world as a whole it has merit. Your example is applying a small group and I get this.
Your original post quoted above seemed to apply the stat to all people in the world, not just young adults with ADHD that are receiving a particular drug. Also, it doesn’t seem like the rate doubled as the article didn’t say the rate was 0.1% and is now 0.21%. Instead the rate for one drug is 0.1% and the rate for a different drug is 0.21%.
If ADHD affect 13% of the population of the world, that is 975,000,000 people out of 7.5 billion. If half of these patients are taking the drug that has the 0.21% psychosis effect like the study, that would be 1,023,750 people experiencing the psychosis. Thus in the global population it would have an effect on 0.014% over an order of magnitude less. But, even this doesn’t really work as ADHD as an adult has much lower diagnosis rates than as a child. Also, 100% of the population of people with ADHD do not get treatment for it.
I hope this clears up the misunderstanding.
Undoubtedly (if they measured a different sample universe, that is).
If you machine something “round” is it truly round? If you use the most precise micrometer you have and you measure its diameter once, and then get a different result (still within spec) when you rotate it 90 degrees and measure it again, is the first measure invalid?
I didn’t intend to mean it could be extrapolated to everyone everywhere, sorry if that was garbled. I was just trying to show what that 0.1% vs 0.21% means in terms of incidence rate. I have edited my post to indicate this.
1 Like
Thanks,
That is totally what had me spinning in my chair with confusion. And when @Lordrook jumped in to corroborate it as the practice of the industry I was dumb founded.
Seems the two of you coming from similar fields and had similar accepted norms for discussion. While I come from marketing where research is often poor and meant to mislead in order to sell.
1 Like
9 out of 10 doctors prefer filtered Camels.
(choice was smoke a cig or play russian roulette)
4 Likes
Yep,
You nailed it. I had to print a coffee ad where a client claimed reverse osmosis filtered water from a spring was better than reverse osmosis water from the city tap. They did a study that claimed this, but I read the study when they asked me to find them quotes to reference and 2 of 3 samples of the spring water registered 0.05% less solids in the water between the tap and the spring and the test had a variability of 0.1%. So the water was identical, but they had a lawyer that claimed they could advertise the difference as a better quality product.